INTRODUCTION

This is the replication package for:

Carie A. Steele, Daniel Pemstein, and Stephen Meserve. Forthcoming. "Democracy Promotion and Electoral Quality: A Disaggregated Analysis." Governance.

This package allows one to reproduce the tables and figures in both
the main text and supplementary materials of the above-cited paper.
The Stata code included in this package was run on a Windows 10 PC
running Stata 16.1 MP. The R code was run on an Ubuntu Linux machine using
version 1.0.9 of the gsynth package. The code in this package is
released under a Creative Commons Attribution-NonCommercial-ShareAlike
4.0 International public license (see LICENSE.txt). The 3rd-party data
included in this package may have other licensing rules; please
consult the appropriate documentation before redistributing those
data.  Use at your own risk.


MANIFEST


***Analysis source code

\Finkelreplication\Finkelreplication.do --- performs all analysis directly using Finkel et al 2007's data.  See analysis instructions, model set 1 for instructions
\Finkelextension\analysis\extension.do --- performs all analysis for the extension models contained in the appendix. see analysis instructions, model set 2 for instructions
\gensynth\analysis.R

***Data construction source code

\Finkelextension\Mergeable aid variables\.do files for creation\creation2001-2019replication.do --- generates USAID dataset that will then merge into vdem, providing 2001-2018 core foreign aid dataset for USAID aid and other US data
\Finkelextension\Mergeable aid variables\.do files for creation\FPPcreation.do --- generates a military aid percentage figure using data from the complete aid dataset
\Finkelextension\Mergeable aid variables\.do files for creation\generateodadata.do --- generates oda dataset for US and total aid from OECD web gui data to merge into vdem
\Finkelextension\Mergeable aid variables\.do files for creation\generateregionaldatastub.do ---  used to generate the regional variables to then merge into v-dem data directly analogous to Finkel's portioning out method
\Finkelextension\Mergeable aid variables\.do files for creation\QoG.do cut down Qog to something manageable and mergeable at the country-year level that creates necessary variables ***NOTE THAT TO SAVE FILE SIZE QOG STANDARD DATASET REFERENCED HERE (qog_std_ts_jan20.dta) IS *NOT* INCLUDED IN THE REPLICATION PACKAGE BUT YOU CAN DOWNLOAD AND SUPPLY IT YOURSELF***
\Finkelextension\Mergeable aid variables\.do files for creation\vdemaiddatamerge.do --- Merges in regional, USAID, and then ODA data.  Also merges in several more datasets for individual independent variables and creates variables extant in the vdem dataset. ***NOTE THAT TO SAVE FILE SIZE VDEM V9 DATASET REFERENCED IN THIS FILE (V-Dem-CY-Full+others-v9.dta) IS *NOT* INCLUDED IN THE REPLICATION PACKAGE BUT YOU CAN DOWNLOAD AND SUPPLY IT YOURSELF***

***Documentation

\readme.txt --- this file, outlines the contents and their use of this replication package
\Finkelreplication\Finkel\Codebook_Phase_2.pdf --- Finkel et al 2007 codebook for reference
\Finkelreplication\Vdem8\V-Dem Codebook v8.pdf --- Vdem v8 codebook for reference
\Finkelextension\Datadictionary_ForeignAidExplorer.pdf --- Foreign Aid Explorer codebook for source data used in extension dataset
\Finkelextension\Updated QoG\qog_std_jan20.pdf --- Codebook for QOG Standard Dataset 2020 used for several variables in the analysi
\Finkelextension\Updated QoG\QoGnote.txt --- placeholder file noting where Qog 2020 dataset should go for full data reconstruction
\Finkelextension\vdem v9\V-Dem Codebook v9.pdf --- Codebook for V-dem v9 data used in extension analysis
\Finkelextension\vdem v9\V-Demnotes --- placeholder file noting where V-dem v9 data should go for full data reconstruction

***Data Description

\Finkelreplication\Finkel\USAID_DG_Programs.dta --- .dta version of SPSS dataset in Finkel et al 2007 replication data folder.
\Finkelreplication\Vdem8\V-Dem-CY+Others-v8.dta --- Vdem v8 .dta data file with only relevant variables included
\Finkelreplication\Vdem8\V-Dem-CY+Others-v8cowchange.dta --- V-Dem v8, generated during data merging, with slight changes to cow data codes for better matching
\Finkelreplication\Vdem8\finkelmerge.dta --- final dataset used in replication and synthetic control
\Finkelextension\finkelreplicationdata20012018.dta --- final dataset used in extension analysis
\Finkelextension\FPPdata20012018 --- military aid variable created by FPPcreation.do
\Finkelextension\oda20002018 --- ODA data variables generated in generateodadata.do
\Finkelextension\ODA20002018allanddgtrimmed.csv --- a trimmed and reorganized data file drawn from a much larger raw dataset taken from the OECD's data on ODA and foreign aid that serves as the source for several control variables.
\Finkelextension\QoGdata20012018.dta --- trimmed QoG data for merging generated by QoG.do
\Finkelextension\us_foreign_aid_completecsv.csv --- full .csv file drawn from USAID's foreign aid explorer.  The source data file for many of the .do's that generate USAID and other foreign aid data figures for analysis.
\Finkelextension\USAID20012018regionsplit.dta --- mergeable regional USAID data generated in generateregionaldatastub.do
\Finkelextension\USAIDdata20022018.dta --- mergable USAID DG and nonDG dtra generated in creation2001-2019replication.do
\gensynth\prepped-gensynth.csv --- .csv file used in synthetic control analysis.

MANIFEST NOTE

For the extension analysis in the appendix, the data of several large, publically available datasets are omitted from this replication folder.  This means to reconstruct the analysis from the core data up, you will need to download the following files:

QOG STANDARD DATASET 2020 (time series): qog_std_ts_jan20.dta
V-Dem v9: V-Dem-CY-Full+Others-v9.dta

ANALYSIS INSTRUCTIONS

The paper, broadly has three sets of models. The files here replicate all three of these sets of models using different sets of .do files for the regression models and X for the synthetic control models.

1) Performing Finkel replication  

The first set of models 1-16 and the descriptive statistics (Tables 1, 2-9) starts from a model based on Finkel et al 2007's specification.   In order to replicate the Finkel et al results using vdem data, consult ReplicationGovernance\Finkelreplication.do.  This file merges in additional data (Vdem8) in order to test whether Finkel's results replicate with micro measures vs. macro measures.  The file ends by identifying the largest changes in the data that is used in section 4.1.

2) Performing Finkel extension

The second set of models 17-28 (Tables 10-12) starts from a model based on Finkel et al 2007 that merges in additional data (USAID, OECD, QoG, vdem) in order to test whether Finkel's results replicate with micro measures vs. macro measures.  If youu choose to completely rebuild the data before performing this analysis, perform the following .do files in the numbered order after downloading and inserting the appropriate vdem v9 and QoG 2020 files which have been omitted for file size reasons.  Also, do not forget to edit the .do files to point at the appropriate folders or they will not function:

1) \Finkelextension\Mergeable aid variables\.do files for creation\creating2001-2018replication.do
2) \Finkelextension\Mergeable aid variables\.do files for creation\generateodadata.do
3) \Finkelextension\Mergeable aid variables\.do files for creation\generateregionaldatastub.do
4) \Finkelextension\Mergeable aid variables\.do files for creation\FPPcreation.do
5) \Finkelextension\Mergeable aid variables\.do files for creation\QoG.do
6) \Finkelextension\Mergeable aid variables\.do files for creation\vdemaiddatamerge.do

If you have recreated the data, or you only choose to replicate the extension analysis, simply run the following .do file to replicate models 17-28 which analyzes Finkelreplicationdata20012018.dta:

\Finkelextension\analysis\extension.do

3) Performing synthetic control models

Run analysis.R in the gensynth subdirectory.
